Back to glossary

KEDA (Kubernetes Event-driven Autoscaling)

A Kubernetes-based Event Driven Autoscaler that allows you to scale any container in Kubernetes based on the number of events needing to be processed.

AbbreviationKEDA
Categorycloud

Overview

While the standard Horizontal Pod Autoscaler (HPA) scales based on CPU and memory, KEDA allows you to scale based on external event sources like Kafka topics, RabbitMQ queues, AWS SQS, or Azure Service Bus. It can even scale deployments down to zero when there are no events.

Key Components

  • Scalers: Connect to external event sources and fetch metrics.
  • Metrics Adapter: Exposes the metrics from scalers to the Kubernetes HPA.
  • Controller: Manages the lifecycle of the 'ScaledObject' and 'ScaledJob' resources.

Benefits

  • Event-Driven Scaling: Scale based on actual workload rather than resource usage.
  • Scale to Zero: Save costs by completely removing pods when they are not needed.
  • Extensible: Supports dozens of different event sources out of the box.